| 遥感影像像元级融合方法与试验评价分析 |
| |
| 引用本文： | 李俊杰,何隆华,戴锦芳,李金莲.遥感影像像元级融合方法与试验评价分析[J].地球信息科学,2008,10(1):128-134. |
| |
| 作者姓名： | 李俊杰 何隆华 戴锦芳 李金莲 |
| |
| 作者单位： | 1.中国资源卫星应用中心, 北京 100073;
2. 中国科学院南京地理与湖泊研究所, 南京 210008;
3. 南京晓庄学院, 南京 210007 |
| |
| 基金项目： | 中国资源卫星应用中心项目，扬州市主要农作物长势监测、中国科学院知识创新工程项目(INF105-SDB-1-22)联合资助 |
| |
| 摘 要： | 不同遥感影像有着各自的优点和局限,遥感影像融合就是为了充分利用它们的优点进行互补得到新的影像。采用PCA、Brovey和小波变换三种方法对SPOT多光谱影像和全色影像进行了像元级的影像融合,小波变换采用的Mallat算法和双正交小波函数,双正交小波函数的优点是小波逆变换时不会导致相位失真,论文对融合结果进行了光谱质量和空间质量两方面的定性和定量的评价。结果表明,小波变换融合影像是三种变换中最优的,同时在小波变换中采用不同的小波基函数的融合结果在光谱质量和空间质量方面也有很大的差异,可根据实际需要选用合适的小波基函数。
|
| 关 键 词： | 遥感影像 小波变换 融合 |
| 收稿时间： | 2007-03-14 |
| 修稿时间： | 7/1/2007 12:00:00 AM |
Analysis of Pixel-level Remote Sensing Image Fusion Methods |
| |
| LI Junjie,HE Longhua,DAI Jinfang,LI Jinlian.Analysis of Pixel-level Remote Sensing Image Fusion Methods[J].Geo-information Science,2008,10(1):128-134. |
| |
| Authors: | LI Junjie HE Longhua DAI Jinfang LI Jinlian |
| |
| Institution: | 1. China Center for Resources Satellite Data and Application, Beijing 100073, China;
2. Nanjing Institute of Geography and Limnology, CAS, Nanjing, Jiangsu 210008, China;
3. Nanjing Xiaozhuang University, Nanjing, Jiangsu 210007, China |
| |
| Abstract: | Different remote sensing images have their own virtues and limits;and image fusion is designed for completely using their virtues to compose new images.Three methods-PCA,Brovey and wavelet transform are used to fuse the SPOT multispectral data and the panchromatic data on the pixel-level.The wavelet transform used the Mallat's algorithm and biorthogonal wavelet function,the virtue of the latter is that the function will not distort the image when the reverse wavelet transform is performed.The fusion results are evaluated both qualitatively and quantitatively in the aspects of spectral and spatial quality;the results show that the best method of the three is wavelet transform.The wavelet transform method makes the fusion image preserve both the spatial quality of the pan image and spectral quality of the multispectral image,while the other two methods make the fusion images to distort the spectral quality of the multispectral image.At the same time,the influences of different wavelet functions used in the Mallat's algorithm on the related fusion image are compared.Different wavelet functions result in differences in reserving the spectral and spatial quality of the fusion image.If we use the biorthogonal wavelet transform to fuse image,we can select the proper wavelet basis function for the practical demand(the compromise between high spatial detail and high spectral quality preserve). |
| |
| Keywords: | remote sensing image wavelet transform fusion |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《地球信息科学》浏览原始摘要信息 |
| 点击此处可从《地球信息科学》下载免费的PDF全文 |
|
